1. Improved unsupervised physics-informed deep learning for intravoxel incoherent motion modeling and evaluation in pancreatic cancer patients
- Author
-
Sebastiano Barbieri, Hans Crezee, Hanneke W. M. van Laarhoven, Misha P. T. Kaandorp, Peter T. While, Remy Klaassen, Aart J. Nederveen, Oliver J. Gurney-Champion, Oncology, CCA - Imaging and biomarkers, AGEM - Amsterdam Gastroenterology Endocrinology Metabolism, Radiotherapy, Radiology and Nuclear Medicine, ACS - Diabetes & metabolism, AMS - Ageing & Vitality, and AMS - Sports
- Subjects
FOS: Computer and information sciences ,Normalization (statistics) ,Computer Science - Machine Learning ,Coefficient of variation ,Bayesian probability ,pancreatic cancer ,Research Articles—Computer Processing and Modeling ,FOS: Physical sciences ,Quantitative Biology - Quantitative Methods ,Least squares ,Standard deviation ,Machine Learning (cs.LG) ,030218 nuclear medicine & medical imaging ,03 medical and health sciences ,Motion ,0302 clinical medicine ,Deep Learning ,Consistency (statistics) ,Statistics ,Humans ,Radiology, Nuclear Medicine and imaging ,Quantitative Methods (q-bio.QM) ,Intravoxel incoherent motion ,intravoxel incoherent motion ,diffusion‐weighted magnetic resonance imaging ,Artificial neural network ,Full Paper ,Physics ,deep neural network ,Reproducibility of Results ,Bayes Theorem ,unsupervised physics-informed deep learning ,Physics - Medical Physics ,IVIM ,Pancreatic Neoplasms ,Diffusion Magnetic Resonance Imaging ,FOS: Biological sciences ,diffusion-weighted magnetic resonance imaging ,Medical Physics (physics.med-ph) ,unsupervised physics‐informed deep learning ,030217 neurology & neurosurgery ,Algorithms - Abstract
Purpose: Earlier work showed that IVIM-NET orig, an unsupervised physics-informed deep neural network, was faster and more accurate than other state-of-the-art intravoxel-incoherent motion (IVIM) fitting approaches to diffusion-weighted imaging (DWI). This study presents a substantially improved version, IVIM-NET optim, and characterizes its superior performance in pancreatic cancer patients. Method: In simulations (signal-to-noise ratio [SNR] = 20), the accuracy, independence, and consistency of IVIM-NET were evaluated for combinations of hyperparameters (fit S0, constraints, network architecture, number of hidden layers, dropout, batch normalization, learning rate), by calculating the normalized root-mean-square error (NRMSE), Spearman’s ρ, and the coefficient of variation (CV NET), respectively. The best performing network, IVIM-NET optim was compared to least squares (LS) and a Bayesian approach at different SNRs. IVIM-NET optim’s performance was evaluated in an independent dataset of 23 patients with pancreatic ductal adenocarcinoma. Fourteen of the patients received no treatment between two repeated scan sessions and nine received chemoradiotherapy between the repeated sessions. Intersession within-subject standard deviations (wSD) and treatment-induced changes were assessed. Results: In simulations (SNR = 20), IVIM-NET optim outperformed IVIM-NET orig in accuracy (NRMSE(D) = 0.177 vs 0.196; NMRSE(f) = 0.220 vs 0.267; NMRSE(D*) = 0.386 vs 0.393), independence (ρ(D*, f) = 0.22 vs 0.74), and consistency (CV NET(D) = 0.013 vs 0.104; CV NET(f) = 0.020 vs 0.054; CV NET(D*) = 0.036 vs 0.110). IVIM-NET optim showed superior performance to the LS and Bayesian approaches at SNRs < 50. In vivo, IVIM-NET optim showed significantly less noisy parameter maps with lower wSD for D and f than the alternatives. In the treated cohort, IVIM-NET optim detected the most individual patients with significant parameter changes compared to day-to-day variations. Conclusion: IVIM-NET optim is recommended for accurate, informative, and consistent IVIM fitting to DWI data.
- Published
- 2021